Develop Reference

bounding boxes on handwritten digits with opencv

I tried the code provided bellow to segment each digit in this image and put a contour around it then crop it out but it's giving me bad results, I'm not sure what I need to change or work on.
The best idea I can think of right now is filtering the 4 largest contours in the image except the image contour itself.
The code I'm working with:
import sys
import numpy as np
import cv2
im = cv2.imread('marks/mark28.png')
im3 = im.copy()
gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray, (5, 5), 0)
thresh = cv2.adaptiveThreshold(blur, 255, 1, 1, 11, 2)
################# Now finding Contours ###################
contours, hierarchy = cv2.findContours(thresh, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
samples = np.empty((0, 100))
responses = []
keys = [i for i in range(48, 58)]
for cnt in contours:
if cv2.contourArea(cnt) > 50:
[x, y, w, h] = cv2.boundingRect(cnt)
if h > 28:
cv2.rectangle(im, (x, y), (x + w, y + h), (0, 0, 255), 2)
roi = thresh[y:y + h, x:x + w]
roismall = cv2.resize(roi, (10, 10))
cv2.imshow('norm', im)
key = cv2.waitKey(0)
if key == 27: # (escape to quit)
sys.exit()
elif key in keys:
responses.append(int(chr(key)))
sample = roismall.reshape((1, 100))
samples = np.append(samples, sample, 0)
responses = np.array(responses, np.float32)
responses = responses.reshape((responses.size, 1))
print
"training complete"
np.savetxt('generalsamples.data', samples)
np.savetxt('generalresponses.data', responses)
I need to change the if condition on height probably but more importantly I need if conditions to get the 4 largest contours on the image. Sadly, I haven't managed to find what I'm supposed to be filtering.
This is the kind of results I get, I'm trying to escape getting those inner contours on the digit "zero"
Unprocessed images as requested: example 1 example 2
All I need is an idea on what I should filter for, don't write code please. Thank you community.

You almost have it. You have multiple bounding rectangles on each digit because you are retrieving every contour (external and internal). You are using cv2.findContours in RETR_LIST mode, which retrieves all the contours, but doesn't create any parent-child relationship. The parent-child relationship is what discriminates between inner (child) and outter (parent) contours, OpenCV calls this "Contour Hierarchy". Check out the docs for an overview of all hierarchy modes. Of particular interest is RETR_EXTERNAL mode. This mode fetches only external contours - so you don't get multiple contours and (by extension) multiple bounding boxes for each digit!
Also, it seems that your images have a red border. This will introduce noise while thresholding the image, and this border might be recognized as the top-level outer contour - thus, every other contour (the children of this parent contour) will not be fetched in RETR_EXTERNAL mode. Fortunately, the border position seems constant and we can eliminate it with a simple flood-fill, which pretty much fills a blob of a target color with a substitute color.
Let's check out the reworked code:
# Imports:
import cv2
import numpy as np
# Set image path
path = "D://opencvImages//"
fileName = "rhWM3.png"
# Read Input image
inputImage = cv2.imread(path+fileName)
# Deep copy for results:
inputImageCopy = inputImage.copy()
# Convert BGR to grayscale:
grayscaleImage = cv2.cvtColor(inputImage, cv2.COLOR_BGR2GRAY)
# Threshold via Otsu:
threshValue, binaryImage = cv2.threshold(grayscaleImage, 0, 255, cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)
The first step is to get the binary image with all the target blobs/contours. This is the result so far:
Notice the border is white. We have to delete this, a simple flood-filling at position (x=0,y=0) with black color will suffice:
# Flood-fill border, seed at (0,0) and use black (0) color:
cv2.floodFill(binaryImage, None, (0, 0), 0)
This is the filled image, no more border!
Now we can retrieve the external, outermost contours in RETR_EXTERNAL mode:
# Get each bounding box
# Find the big contours/blobs on the filtered image:
contours, hierarchy = cv2.findContours(binaryImage, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
Notice you also get each contour's hierarchy as second return value. This is useful if you want to check out if the current contour is a parent or a child. Alright, let's loop through the contours and get their bounding boxes. If you want to ignore contours below a minimum area threshold, you can also implement an area filter:
# Look for the outer bounding boxes (no children):
for _, c in enumerate(contours):
# Get the bounding rectangle of the current contour:
boundRect = cv2.boundingRect(c)
# Get the bounding rectangle data:
rectX = boundRect[0]
rectY = boundRect[1]
rectWidth = boundRect[2]
rectHeight = boundRect[3]
# Estimate the bounding rect area:
rectArea = rectWidth * rectHeight
# Set a min area threshold
minArea = 10
# Filter blobs by area:
if rectArea > minArea:
# Draw bounding box:
color = (0, 255, 0)
cv2.rectangle(inputImageCopy, (int(rectX), int(rectY)),
(int(rectX + rectWidth), int(rectY + rectHeight)), color, 2)
cv2.imshow("Bounding Boxes", inputImageCopy)
# Crop bounding box:
currentCrop = inputImage[rectY:rectY+rectHeight,rectX:rectX+rectWidth]
cv2.imshow("Current Crop", currentCrop)
cv2.waitKey(0)
The last three lines of the above snippet crop and show the current digit. This is the result of detected bounding boxes for both of your images (the bounding boxes are colored in green, the red border is part of the input images):

Bounding Box selecting entire image instead of polygons

I am trying to draw bounding boxes around polygon images.I orginally marked the images and created the polygon masks using a json file.
This is my code:
I use the json file to keep the names of the files constant.
import cv2
import numpy as np
import json
import matplotlib.pyplot as plt
jsonFile ='/directory..../.json'
with open(jsonFile) as file:
annotations = json.load(file)
for key in annotations:
regions = annotations[key]['regions']
for region in regions:
print(annotations[key]['filename'],"\n")
image = cv2.imread('/directory to mask images.png' + annotations[key]['filename'])
original = image.copy()
gray=cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)
thresh = cv2.threshold(gray, 0,255,cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]
# Find contours, obtain bounding box, extract and save ROI
ROI_number = 0
cnts = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
for c in cnts:
x,y,w,h = cv2.boundingRect(c)
cv2.rectangle(image, (x, y), (x + w, y + h), (36,255,12), 2)
ROI = original[y:y+h, x:x+w]
cv2.imwrite('ROI_{}.png'.format(annotations[key]['filename']), ROI)
ROI_number += 1
cv2.imshow('image', image)
cv2.waitKey()
My problem is instead of drawing bounding boxes over the polygons in the picture, I am just outlining the border of the full image.
If it matters, the image is black and the polygon masks are white.

I think the problem might come from the threshold operation. If the original image is black with white polygons, then the cv2.THRESH_BINARY_INV will transform it into a white image with black polygons. Since the contour function finds the polygons surrounding the white space, the resulting polygon will surround the whole image. In order to fix this, just use cv2.THRESH_BINARY instead of cv2.THRESH_BINARY_INV.

How to crop an image along major axis of the blob/contour with a rectangular bounding box using Python

I would like to crop an image which has a hand drawn highlighted area in orange as shown below,
The result should be a cropped image along the major axis of the blob or contour with a rectangular bounding box, as shown below,
Here's what i have tried,
import numpy as np
import cv2
# load the image
image = cv2.imread("frame50.jpg", 1)
#color boundaries [B, G, R]
lower = [0, 3, 30]
upper = [30, 117, 253]
# create NumPy arrays from the boundaries
lower = np.array(lower, dtype="uint8")
upper = np.array(upper, dtype="uint8")
# find the colors within the specified boundaries and apply
# the mask
mask = cv2.inRange(image, lower, upper)
output = cv2.bitwise_and(image, image, mask=mask)
ret,thresh = cv2.threshold(mask, 50, 255, 0)
if (int(cv2.__version__[0]) > 3):
contours, hierarchy = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
else:
im2, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
if len(contours) != 0:
# find the biggest countour (c) by the area
c = max(contours, key = cv2.contourArea)
x,y,w,h = cv2.boundingRect(c)
ROI = image[y:y+h, x:x+w]
cv2.imshow('ROI',ROI)
cv2.imwrite('ROI.png',ROI)
cv2.waitKey(0)
This does not seem to work most of the time.For some images, the following happens,
I would like to know if there is better way to go about this or how i can fix what i have right now.Note that the highlighted area is hand drawn and can be of any shape but it is closed and not left open and the colour of the highlight is that shade of orange itself in all cases.
And is there a way to only retain content inside the circle and blackout everything outside it?
EDIT1:
I was able to fix the wrong clipping by varying the threshold more. But my main query now is: is there a way to only retain content inside the circle and blackout everything outside it? I can see the mask as show below,
How do I fill this mask and retain content inside the circle and blackout everything outside it, with the same rectangular bounding box?

Have you tried
image[x:x+w, y:y+h]
And could you check bbox with below code
cv2.rectangle(thresh,(x,y),(x+w,y+h),(255,0,0),2)

First of all, it is always better to use an HSV image instead of BGR image for masking(extracting a color). You can do this by the following code.
HSV_Image = cv2.cvtColor(Image, cv2.COLOR_BGR2HSV)
ThreshImage = cv2.inRange(HSV_Image, np.array([0, 28, 191]), np.array([24, 255, 255]))
The range numbers here are found for orange color in this case.
Image is the input image and ThreshImage is the output image with the orange region colored as white and everything else as black.
Now finding the contour in ThreshImage with cv2.RETR_EXTERNAL flag will give only one contour that is the outer boundary of the orange region.
Contours, Hierarchy = cv2.findContours(ThreshImage, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
To crop the orange region:
BoundingRect = cv2.boundingRect(Contours[0])
(x, y, w, h) = BoundingRect
CroppedImage = Image[y:y+h, x:x+w].copy()
"CroppedImage" will store the cropped orange region as desired.
To get contents of only inside the contour:
Bitwise AND operation will be useful here as we already have detected the contour.
First, we have to create a black image of shape the same as that of input image and draw the contour with white color filled in it.
ContourFilledImage = np.zeros(Image.shape, dtype=np.uint8)
cv2.drawContours(ContourFilledImage, Contours, -1, (255, 255, 255), -1)
Now perform a bitwise AND operation on Input Image and "ContourFilledImage"
OnlyInnerData = cv2.bitwise_and(ContourFilledImage, Image)
"OnlyInnerData" image is the desired output image having only the content of inside the circle.

Extracting and saving characters from an image

I am following up from this post: How to extract only characters from image?
This solution works perfectly for me (with some tweaking) for its desired purpose. However, I am attempting to take it 1 step further by saving each of the characters. So in the example of this post, I wish to save the characters K, N, and M as their own individual images. I attempted iterating through the nested if loop with a cv2.imwrite function with the rect object although the final output is 7 images containing the overall image with just an additional rectangle to highlight the next contour each time.

Here's a simple approach:
Obtain binary image. Load image, grayscale, Otsu's threshold
Extract ROIs. Find contours
and sort from left-to-right to ensure we have the contours in the correct order with imutils.contours.sort_contours. We filter using contour area then extract and save each ROI using Numpy slicing.
Input
Binary image
Detected characters highlighted in green
Extracted ROIs
Code
import cv2
from imutils import contours
# Load image, grayscale, Otsu's threshold
image = cv2.imread('1.png')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY_INV)[1]
# Find contours, sort from left-to-right, then crop
cnts = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
cnts, _ = contours.sort_contours(cnts, method="left-to-right")
# Filter using contour area and extract ROI
ROI_number = 0
for c in cnts:
area = cv2.contourArea(c)
if area > 10:
x,y,w,h = cv2.boundingRect(c)
ROI = image[y:y+h, x:x+w]
cv2.imwrite('ROI_{}.png'.format(ROI_number), ROI)
cv2.rectangle(image, (x, y), (x + w, y + h), (36,255,12), 2)
ROI_number += 1
cv2.imshow('thresh', thresh)
cv2.imshow('image', image)
cv2.waitKey()

How to remove a contour inside contour in Python OpenCV?

OpenCV in Python provides the following code:
regions, hierarchy = cv2.findContours(binary_image, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
for region in regions:
x, y, w, h = cv2.boundingRect(region)
cv2.rectangle(image, (x, y), (x + w, y + h), (0, 255, 0), 1)
This gives some contours within contour. How to remove them in Python?

For that, you should take a look at this tutorial on how to use the hierarchy object returned by the method findContours .
The main point is that you should use cv2.RETR_TREE instead of cv2.RETR_LIST to get parent/child relationships between your clusters:
regions, hierarchy = cv2.findContours(binary_image, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
Then you can check whether a contour with index i is inside another by checking if hierarchy[0,i,3] equals -1 or not. If it is different from -1, then your contour is inside another.

In order to remove the contours inside a contour:
shapes, hierarchy = cv2.findContours(image=image, mode=cv2.RETR_EXTERNAL, method=cv2.CHAIN_APPROX_SIMPLE)
However, in some cases you may observe that a big contour is formed on the whole image, and applying the above returns you that one big contour.
In order to avoid this, try inverting the image:
image = cv2.imread("Image Path")
image = 255 - image
shapes, hierarchy = cv2.findContours(image=image, mode=cv2.RETR_EXTERNAL, method=cv2.CHAIN_APPROX_SIMPLE)
This will give you the desired result.
UPDATE:
The reason why hierarchy does not work if a big bounding box is approximated on the whole image is that the output of hierarchy[0,iteration,3] is -1 only for the one bounding box drawn on the whole image, as all other bounding boxes are inside this big bounding box, and hierarchy[0,iteration,3] is not equal to -1 for any of them. Thus, inverting the image will be required in order to comply with the following:
In OpenCV, finding contours is like finding white object from black background. So remember, object to be found should be white and background should be black.
However, as pointed out by #Jeru, this is not a generalized solution and one must visualize the image before inverting it.
Consider this image:
Running
shapes, hierarchy = cv2.findContours(image=image, mode=cv2.RETR_TREE, method=cv2.CHAIN_APPROX_SIMPLE)
results in
Now, only displaying the contour with hierarchy[0,iteration,3] = -1 results in
which is not correct. If we want to obtain the rectangle containing the shapes and the text shapes, we can do
image = 255 - image
shapes, hierarchy = cv2.findContours(image=thresh, mode=cv2.RETR_EXTERNAL, method=cv2.CHAIN_APPROX_SIMPLE)
In this case we get:
Code:
import cv2
from easyocr import Reader
import math
shape_number = 2
image = cv2.imread("Image Path")
deep_copy = image.copy()
image_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(image_gray, 210, 255, cv2.THRESH_BINARY)
thresh = 255 - thresh
shapes, hierarchy = cv2.findContours(image=thresh, mode=cv2.RETR_EXTERNAL, method=cv2.CHAIN_APPROX_SIMPLE)
cv2.drawContours(image=deep_copy, contours=shapes, contourIdx=-1, color=(0, 255, 0), thickness=2, lineType=cv2.LINE_AA)
for iteration, shape in enumerate(shapes):
if hierarchy[0,iteration,3] == -1:
print(hierarchy[0,iteration,3])
print(iteration)
cv2.imshow('Shapes', deep_copy)
cv2.waitKey(0)
cv2.destroyAllWindows()

img_output, contours, hierarchy = cv2.findContours(blank_image_firstImage, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
This removes the child contour